MemorizedMCP / rules

---
description:
globs:
alwaysApply: true
---

# MCP Server AI Agent Usage Guide

## Table of Contents

- [MCP Server AI Agent Usage Guide](#mcp-server-ai-agent-usage-guide)
  - [Table of Contents](#table-of-contents)
  - [Sequential Thinking MCP Server](#sequential-thinking-mcp-server)
    - [Purpose](#purpose)
    - [Supported Operations](#supported-operations)
      - [Additional Inputs](#additional-inputs)
  - [FileSystem MCP Server](#filesystem-mcp-server)
    - [Purpose](#purpose-1)
    - [Supported Operations](#supported-operations-1)
  - [Fetch MCP Server](#fetch-mcp-server)
    - [Purpose](#purpose-2)
    - [Supported Operations](#supported-operations-2)
  - [Git MCP Server](#git-mcp-server)
    - [Purpose](#purpose-3)
    - [Supported Operations](#supported-operations-3)
  - [ShadCN-UI MCP Server](#shadcn-ui-mcp-server)
    - [Purpose](#purpose-4)
    - [Supported Operations](#supported-operations-4)
  - [React MCP Server](#react-mcp-server)
    - [Purpose](#purpose-5)
    - [Supported Operations](#supported-operations-5)
  - [NodeJS MCP Server](#nodejs-mcp-server)
    - [Purpose](#purpose-6)
    - [Supported Operations](#supported-operations-6)
  - [Chrome DevTools MCP Server](#chrome-devtools-mcp-server)
    - [Purpose](#purpose-7)
    - [Supported Operations](#supported-operations-7)
  - [Grep MCP Server](#grep-mcp-server)
    - [Purpose](#purpose-8)
    - [Supported Operations](#supported-operations-8)
  - [MemorizedMCP Server](#memorizedmcp-server)
    - [Purpose](#purpose-9)
    - [Supported Operations](#supported-operations-9)

## Sequential Thinking MCP Server

### Purpose

Helps organize thoughts, break down problems, and build reasoning step by step.

### Supported Operations

- **sequential_thought:** Submit a reasoning step or thought.
    - Example:

```json
{ "thought": "Step 1: Identify the problem" }
```

- **progress_status:** Track progress within a sequence of thoughts.
    - Example:

```json
{}
```

- **branch_thought:** Create a branched path in reasoning from a previous step.
    - Example:

```json
{ "branchFromThought": 2, "branchId": "alternative1" }
```

- **revise_thought:** Update or revise a previous thought.
    - Example:

```json
{ "revisesThought": 3, "thought": "Updated consideration for step 3" }
```


#### Additional Inputs

- `thought`: The current reasoning step.
- `nextThoughtNeeded`: Boolean for more steps.
- `thoughtNumber`: Current step index.
- `totalThoughts`: Planned total steps.
- Branching supports `branchFromThought` and `branchId`.


## FileSystem MCP Server

### Purpose

Allows safe interaction with files and directories within allowed paths.

### Supported Operations

- **read_file:** Read the contents of a file.
    - Example:

```json
{ "path": "/projects/myfile.txt" }
```

- **read_multiple_files:** Read several files at once.
    - Example:

```json
{ "paths": ["/projects/a.txt", "/projects/b.txt"] }
```

- **write_file:** Overwrite or create a file with desired content.
    - Example:

```json
{ "path": "/projects/output.txt", "content": "Hello World" }
```

- **edit_file:** Apply specified edits to a file (use dryRun for safety).
    - Example:

```json
{ "path": "/projects/config.json", "edits": [ { "oldText": "foo", "newText": "bar" } ], "dryRun": true }
```

- **create_directory:** Make a new directory, parents included.
    - Example:

```json
{ "path": "/projects/newdir" }
```

- **list_directory:** Show files and folders inside a directory.
    - Example:

```json
{ "path": "/projects" }
```

- **move_file:** Move or rename files and directories.
    - Example:

```json
{ "source": "/projects/old.txt", "destination": "/projects/new.txt" }
```

- **search_files:** Recursively find files by pattern.
    - Example:

```json
{ "path": "/projects", "pattern": "*.md" }
```

- **get_file_info:** Get details such as size and timestamps.
    - Example:

```json
{ "path": "/projects/readme.md" }
```

- **list_allowed_directories:** Show which directories are accessible.
    - Example:

```json
{}
```


## Fetch MCP Server

### Purpose

Fetches web content from URLs, returning plain text or markdown.

### Supported Operations

- **fetch:** Retrieve web page content as markdown or raw data.
    - Example:

```json
{
  "url": "https://example.com",
  "max_length": 5000,
  "start_index": 0,
  "raw": false
}
```

    - `url`: Required.
    - `max_length`: Optional limit (default 5000 characters).
    - `start_index`: Optional, starts at a specific index.
    - `raw`: Optional, set to true for raw HTML/text.


## Git MCP Server

### Purpose

Allows automated version control on repositories including cloning, commits, status, and logs.

### Supported Operations

- **git_clone:** Clone a repository by URL to a destination.
    - Example:

```json
{ "repo_url": "https://github.com/user/repo.git", "destination": "/projects/repo" }
```

- **git_commit:** Commit staged changes with a message.
    - Example:

```json
{ "message": "Fix issue #123" }
```

- **git_push:** Push committed changes upstream.
    - Example:

```json
{}
```

- **git_pull:** Pull the latest changes from the remote.
    - Example:

```json
{}
```

- **git_status:** Check for changed or untracked files.
    - Example:

```json
{}
```

- **git_log:** View recent commit history, optionally limit results.
    - Example:

```json
{ "limit": 10 }
```


## ShadCN-UI MCP Server

### Purpose

Lets you access shadcn/ui v4 React components, code, demos, blocks, and metadata.

### Supported Operations

- **get_component:** Get a component’s source code.
    - Example:

```json
{ "componentName": "button" }
```

- **get_component_demo:** See component usage examples.
    - Example:

```json
{ "componentName": "button" }
```

- **list_components:** Show all available components.
    - Example:

```json
{}
```

- **get_component_metadata:** Fetch dependencies or config for a component.
    - Example:

```json
{ "componentName": "button" }
```

- **get_block:** Fetch a full UI block (like a signup or dashboard set).
    - Example:

```json
{ "blockName": "dashboard-01" }
```

- **list_blocks:** List every block, grouped by type.
    - Example:

```json
{}
```

- **get_directory_structure:** Show the component library directory tree.
    - Example:

```json
{}
```


## React MCP Server

### Purpose

Automates full React project workflows: creation, file management, running, and process control.

### Supported Operations

- **create-react-app:** Make a new app, optionally with a template.
    - Example:

```json
{ "name": "my-app", "template": "typescript", "directory": "/projects" }
```

- **run-react-app:** Start the development server for an app.
    - Example:

```json
{ "projectPath": "/projects/my-app" }
```

- **run-command:** Run any shell command in a directory.
    - Example:

```json
{ "command": "npm run build", "directory": "/projects/my-app" }
```

- **get-process-output:** Get output logs from a running process.
    - Example:

```json
{ "processId": "123456" }
```

- **stop-process:** Stop a process by its ID.
    - Example:

```json
{ "processId": "123456" }
```

- **list-processes:** Show all current running processes.
    - Example:

```json
{}
```

- **edit-file:** Write or update file content.
    - Example:

```json
{ "filePath": "/projects/my-app/src/App.js", "content": "// new code" }
```

- **read-file:** Read a specific file’s contents.
    - Example:

```json
{ "filePath": "/projects/my-app/src/App.js" }
```

- **install-package:** Install an npm package, optional dev flag.
    - Example:

```json
{ "packageName": "react-router-dom", "directory": "/projects/my-app", "dev": true }
```

- **check-installation-status:** Review package installation state.
    - Example:

```json
{ "processId": "987654" }
```


## NodeJS MCP Server

### Purpose

Controls and automates execution of Node.js scripts, npm scripts, and direct eval code.

### Supported Operations

- **run-node-script:** Run a script file with arguments and input.
    - Example:

```json
{
  "scriptPath": "/scripts/my-script.js",
  "nodeArgs": [],
  "args": ["arg1", "arg2"],
  "stdin": "input data",
  "cwd": "/projects"
}
```

- **run-npm-script:** Execute scripts from a package.json in a directory.
    - Example:

```json
{
  "packageDir": "/projects/my-app",
  "scriptName": "start",
  "args": [],
  "stdin": ""
}
```

- **run-node-eval:** Evaluate JavaScript code within Node.js.
    - Example:

```json
{
  "code": "console.log('Hello world')",
  "evalDirectory": "/projects/my-app",
  "stdin": ""
}
```

- **node-version:** Get information about the Node.js runtime.
    - Example:

```json
"node-version://info"
```

- **npm-scripts:** List possible npm scripts for a directory.
    - Example:

```json
"npm-scripts:///projects/my-app"
```


## Chrome DevTools MCP Server

### Purpose

Automates Chrome debugging and page interaction via the DevTools Protocol. Enables navigation, DOM actions, network inspection, performance tracing, and screenshots.

### Supported Operations

- **list_console_messages:** List all console messages for the currently selected page.
  - Example:

```json
{}
```

- **emulate_cpu:** Emulate CPU throttling.
  - Inputs: `throttlingRate` (number, 1-20)
  - Example:

```json
{ "throttlingRate": 4 }
```

- **emulate_network:** Emulate network conditions.
  - Inputs: `throttlingOption` ("No emulation" | "Slow 3G" | "Fast 3G" | "Slow 4G" | "Fast 4G")
  - Example:

```json
{ "throttlingOption": "Fast 3G" }
```

- **click:** Click an element.
  - Inputs: `uid` (string), `dblClick` (optional boolean)
  - Example:

```json
{ "uid": "el_123" }
```

- **drag:** Drag an element onto another.
  - Inputs: `from_uid` (string), `to_uid` (string)
  - Example:

```json
{ "from_uid": "el_item", "to_uid": "el_dropzone" }
```

- **fill:** Type into an input/textarea or select an option.
  - Inputs: `uid` (string), `value` (string)
  - Example:

```json
{ "uid": "el_email", "value": "user@example.com" }
```

- **fill_form:** Fill multiple elements at once.
  - Inputs: `elements`: [{ `uid`, `value` }]
  - Example:

```json
{ "elements": [ { "uid": "el_email", "value": "user@example.com" }, { "uid": "el_pwd", "value": "secret" } ] }
```

- **hover:** Hover over an element.
  - Inputs: `uid` (string)
  - Example:

```json
{ "uid": "el_menu" }
```

- **upload_file:** Upload a file via an input.
  - Inputs: `uid` (string), `filePath` (string)
  - Example:

```json
{ "uid": "el_file", "filePath": "C:/path/file.png" }
```

- **get_network_request:** Get a specific network request by URL.
  - Inputs: `url` (string)

```json
{ "url": "https://example.com/api/data" }
```

- **list_network_requests:** List network requests with optional filtering.
  - Inputs: `pageSize` (optional number), `pageIdx` (optional number), `resourceTypes` (optional array)
  - Example:

```json
{ "pageSize": 50, "resourceTypes": ["xhr", "fetch"] }
```

- **close_page:** Close a page by index (cannot close the last open page).
  - Inputs: `pageIdx` (number)
  - Example:

```json
{ "pageIdx": 1 }
```

- **handle_dialog:** Accept/dismiss browser dialogs.
  - Inputs: `action` ("accept" | "dismiss"), `promptText` (optional)
  - Example:

```json
{ "action": "accept" }
```

- **list_pages:** List open pages.
  - Example:

```json
{}
```

- **navigate_page:** Navigate the selected page to a URL.
  - Inputs: `url` (string)
  - Example:

```json
{ "url": "https://example.com" }
```

- **navigate_page_history:** Go back or forward in history.
  - Inputs: `navigate` ("back" | "forward")
  - Example:

```json
{ "navigate": "back" }
```

- **new_page:** Create a new page.
  - Inputs: `url` (string)
  - Example:

```json
{ "url": "https://example.com" }
```

- **resize_page:** Resize the window of the selected page.
  - Inputs: `width` (number), `height` (number)
  - Example:

```json
{ "width": 1280, "height": 800 }
```

- **select_page:** Select a page by index.
  - Inputs: `pageIdx` (number)
  - Example:

```json
{ "pageIdx": 0 }
```

- **performance_analyze_insight:** Get details on a specific performance insight.
  - Inputs: `insightName` (string)
  - Example:

```json
{ "insightName": "DocumentLatency" }
```

- **performance_start_trace:** Start a performance trace, optionally auto-reload/stop.
  - Inputs: `reload` (boolean), `autoStop` (boolean)
  - Example:

```json
{ "reload": true, "autoStop": true }
```

- **performance_stop_trace:** Stop active trace.
  - Example:

```json
{}
```

- **take_screenshot:** Screenshot page or element.
  - Inputs: `format` ("png" | "jpeg"), `uid` (optional string), `fullPage` (optional boolean)
  - Example:

```json
{ "format": "png", "fullPage": true }
```

- **evaluate_script:** Evaluate JS on the selected page.
  - Inputs: `function` (stringified function), `args` (optional array of element refs)
  - Example:

```json
{ "function": "() => document.title" }
```

- **take_snapshot:** Capture a text snapshot of the page to obtain element UIDs.
  - Example:

```json
{}
```

- **wait_for:** Wait for text to appear on the page.
  - Inputs: `text` (string)
  - Example:

```json
{ "text": "Welcome" }
```


## Grep MCP Server

### Purpose

Searches over a large index of public GitHub repositories. Returns code snippets matching patterns with filters for language, repository, and file path.

### Supported Operations

- **grep-search:** Execute a code search.
  - Inputs:
    - `query` (string): literal code or regex pattern
    - `matchCase` (optional boolean)
    - `matchWholeWords` (optional boolean)
    - `useRegexp` (optional boolean)
    - `repo` (optional string)
    - `path` (optional string)
    - `language` (optional string[])
  - Example (regex across languages):

```json
{
  "query": "(?s)server\\.tool.*catch",
  "language": ["TypeScript", "JavaScript"],
  "useRegexp": true
}
```


## MemorizedMCP Server

### Purpose

Provides a hybrid memory system (knowledge graph + embeddings + full‑text + documentary memory) for agents to store and retrieve information with rich relationships.

### Supported Operations

- **system.status:** Return health and runtime metrics (uptime, indices, storage, metrics, memory, health).
  - Example:

```json
{}
```

- **system.cleanup:** Maintenance cleanup.
  - Inputs: `reindex` (boolean), `compact` (boolean)
  - Example:

```json
{ "reindex": false, "compact": true }
```

- **system.backup:** Create a backup snapshot.
  - Inputs: `destination` (optional string), `includeIndices` (boolean)
  - Example:

```json
{ "destination": "./backups", "includeIndices": true }
```

- **system.restore:** Restore from snapshot.
  - Inputs: `source` (string), `includeIndices` (boolean)
  - Example:

```json
{ "source": "./backups/2025-09-27", "includeIndices": true }
```

- **document.store:** Ingest PDF/Markdown/Text; parse and index.
  - Inputs: `path` (optional string), `mime` (optional "pdf"|"md"|"txt"), `content` (optional string), `metadata` (object)
  - Example:

```json
{ "mime": "md", "content": "# Title\nHello" }
```

- **document.retrieve:** Retrieve document by id/hash/path.
  - Inputs: one of `id` | `hash` | `path`
  - Example:

```json
{ "id": "DOC_123" }
```

- **document.analyze:** Get key concepts, entities, summary, related docs.
  - Inputs: `id` (string)
  - Example:

```json
{ "id": "DOC_123" }
```

- **document.refs_for_memory:** List document refs associated with a memory.
  - Inputs: `id` (memory id)
  - Example:

```json
{ "id": "MEM_123" }
```

- **document.refs_for_document:** List memories referencing a document.
  - Inputs: `id` (document id)
  - Example:

```json
{ "id": "DOC_123" }
```

- **document.validate_refs:** Validate/correct memory↔document references.
  - Inputs: `fix` (boolean)
  - Example:

```json
{ "fix": true }
```

- **kg.list_entities:** List top entities by mention count.
  - Inputs: `limit` (number)
  - Example:

```json
{ "limit": 50 }
```

- **kg.get_entity:** Get detailed information about an entity.
  - Inputs: `entity` (string)
  - Example:

```json
{ "entity": "Python" }
```

- **kg.create_entity:** Create or ensure an entity node exists.
  - Inputs: `entity` (string)
  - Example:

```json
{ "entity": "Python" }
```

- **kg.create_relation:** Create a relation between two nodes.
  - Inputs: `src` (string), `dst` (string), `relation` (string, optional)
  - Example:

```json
{ "src": "Entity::Python", "dst": "Document::abc123", "relation": "MENTIONS" }
```

- **kg.search_nodes:** Search nodes by type and pattern.
  - Inputs: `type` (string, optional), `pattern` (string, optional), `limit` (number)
  - Example:

```json
{ "type": "Entity", "pattern": "python", "limit": 10 }
```

- **kg.read_graph:** Get graph snapshot with configurable limit.
  - Inputs: `limit` (number)
  - Example:

```json
{ "limit": 100 }
```

- **kg.tag_entity:** Add tags to an entity.
  - Inputs: `entity` (string), `tags` (array of strings)
  - Example:

```json
{ "entity": "Python", "tags": ["language", "backend"] }
```

- **kg.get_tags:** Get all tags or entities by a specific tag.
  - Inputs: `tag` (string, optional)
  - Example:

```json
{ "tag": "language" }
```

- **kg.remove_tag:** Remove tags from an entity.
  - Inputs: `entity` (string), `tags` (array of strings)
  - Example:

```json
{ "entity": "Python", "tags": ["backend"] }
```

- **kg.delete_entity:** Delete an entity and its edges.
  - Inputs: `entity` (string)
  - Example:

```json
{ "entity": "Python" }
```

- **kg.delete_relation:** Delete a specific relation.
  - Inputs: `src` (string), `dst` (string), `relation` (string, optional)
  - Example:

```json
{ "src": "Entity::Python", "dst": "Document::abc123", "relation": "MENTIONS" }
```

- **memory.add:** Add a memory; classify STM/LTM; optionally link docs.
  - Inputs: `content` (string), `metadata` (object, optional), `references` (array of { `docId` (string), `chunkId` (optional string), `score` (optional number) }), `layer_hint` ("STM"|"LTM"), `session_id` (optional string), `episode_id` (optional string)
  - Example:

```json
{
  "content": "User prefers dark mode",
  "references": [ { "docId": "DOC_123", "chunkId": "2" } ],
  "layer_hint": "STM"
}
```

- **memory.search:** Hybrid search with temporal filters.
  - Inputs: `q` (string), `limit` (number), `layer` (optional "STM"|"LTM"), `episode` (optional string), `from` (epoch ms), `to` (epoch ms)
  - Example:

```json
{ "q": "dark mode", "limit": 20 }
```

- **memory.update:** Update memory content/metadata.
  - Inputs: `id` (string), `content` (optional string), `metadata` (optional object)
  - Example:

```json
{ "id": "MEM_123", "content": "Updated preference" }
```

- **memory.delete:** Delete a memory (optionally back up first).
  - Inputs: `id` (string), `backup` (boolean)
  - Example:

```json
{ "id": "MEM_123", "backup": true }
```

- **advanced.consolidate:** Promote STM → LTM, deduplicate salient memories.
  - Inputs: `dryRun` (boolean), `limit` (number)
  - Example:

```json
{ "dryRun": false, "limit": 100 }
```

- **advanced.analyze_patterns:** Mine recurring concepts and trends.
  - Inputs: `window` ({ `from`: number, `to`: number }), `minSupport` (number)
  - Example:

```json
{ "window": { "from": 0, "to": 4102444800000 }, "minSupport": 2 }
```

- **advanced.reindex:** Rebuild indices.
  - Inputs: `vector` (boolean), `text` (boolean), `graph` (boolean)
  - Example:

```json
{ "vector": true, "text": true, "graph": true }
```

- **advanced.trends:** Time-bucketed memory counts.
  - Inputs: `from` (number), `to` (number), `buckets` (number)
  - Example:

```json
{ "from": 0, "to": 4102444800000, "buckets": 10 }
```

- **advanced.clusters:** Document clusters.
  - Inputs: `{}`
  - Example:

```json
{}
```

- **advanced.relationships:** Relationship group counts.
  - Inputs: `{}`
  - Example:

```json
{}
```

- **advanced.effectiveness:** Memory effectiveness scores.
  - Inputs: `{}`
  - Example:

```json
{}
```